nemoclaw-maintainer-analyze-ci-performance
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local TypeScript scripts via Node.js, which in turn use standard system utilities (
bash,gh,unzip,awk,dd) for data processing. It implements a custom quoting function to safely escape all shell arguments, preventing command injection from user-provided inputs like repository names. - [EXTERNAL_DOWNLOADS]: The skill downloads CI artifacts (ZIP files) from GitHub using the authenticated
ghCLI. It enforces a strict 25MB limit on compressed data and includes built-in safeguards against resource exhaustion by bounding stream reads. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of Vitest JSON reports. It mitigates potential injection or confusion attacks by applying a strict internal schema, validating ZIP entry paths to prevent traversal, and restricting output to numerical timing statistics.
- [CREDENTIALS_UNSAFE]: The skill utilizes the user's existing GitHub CLI session for authentication. It includes a comprehensive redaction utility (
redactDiagnostic) that automatically filters out tokens, authorization headers, and sensitive local file paths (like home directories) from any error messages or diagnostic output.
Audit Metadata